We have all heard about the long-tail of search marketing. And odds are that even if you have not intentionally implemented such a strategy, you are to some extent utilising broad or phrase match keywords, thus capturing a long tail of varyingly relevant searches.
Long-tail keywords are phrases that are not often searched for individually, but in aggregate can create a sizeable opportunity. And with increased user sophistication and complexity of queries, the opportunity grows larger.
According to an Experian Hitwise US report (October 2009), about 60% of searches consisted of only one to three words. What is more intriguing though, is the increasing proportion of longer searches, constructed by five or more words.
Another eye-opening statistic is from Google, stating that 20-25% of all searches are brand new, ergo having never before been searched for.
Simplistically, there are two approaches to long-tail paid search: source thousands of relevant keywords in an effort to capture them all as exact matches, or rely on the broad and phrase matches to trigger the tail searches. The optimal solution is rarely one or the other, but a mix of tactics. The perfect sauce depends on available tools and processes.
Keyword length and unique queries
To find out what number of search phrases have resulted in a click, simply pull a summary “search query” report from Google, or similar report that shows actual search queries, de-duplicate and count the number of rows.
A recent campaign analysis in the software sector revealed that the total 1000 clicks resulted from over 750 unique search queries, and no keyword clicked more than twice; a proper long-tail that drove volume traffic.
For analysing the keyword length, Excel is often the tool of choice. The following function will count the number of words (assuming the query is in column A with headers) by using the empty spaces to separate words, while removing leading and trailing spaces:
A good visual representation of how the long-tail works is a line chart plotting the keyword length and the number of impressions. These will typically have an inverted relationship.
So what do these two pieces of statistics actually tell you?
Not much in isolation. However, they are all useful in conjunction with analysing the quality of search phrases, comparing the patterns with similar campaigns for benchmarking and ultimately understanding how well the current campaign structure is corresponding to actual user behaviour.